As a kid, I would copy down—by hand—the entirety of encyclopedia entries about animals. In fifth grade, I was regularly running natural history experiments (see photo). By senior year of high school, I was determined to earn, “PhDs in Biology and Ecology” (see embarrassing photo).

That goal might sound ambitious, but not outlandish to most folks, especially those currently in academia. But where I come from, it was in league with aspirations to become an NFL footballer, a rockstar, or lottery jackpot winner—life outcomes that, while theoretically possible, were implausible to the point of fantasy.

I come from a low-income family. Neither of my parents went to college. So, a college education, let alone graduate school, was not an expectation. Much the opposite. The fact that I went to college at all was a universal alignment of serendipity. If it hadn’t been for some friends explaining the process, a recycled plastic frisbee, and my dad’s meager life insurance after his overdose in my junior year, I certainly would not have gone to college.

Nevertheless, I managed to make that dream of a doctorate—from Yale, no less—come true. I even had my dream post-doc lined up with an amazing set of mentors and collaborators.

Then I decided to walk away.

That decision was the hardest, but simultaneously, the most obvious decision point of my entire life. Judging from the whinging about a lack of post-docs on science Twitter, it seems like others are taking a long hard look at the prospects of life in the academy and deciding to opt out as well.

For others at the decision point, here are some things to consider:

Hands-down, academics have the greatest disjunct between an overly inflated concept of self-worth yet accept horribly low monetary valuation of their worth. The average doctorate stipend in the US is somewhere between $15k and $30k. That’s less than California’s minimum wage going to lots of folks who already have Master’s degrees.

This paltry payout is enabled because most academics are heavily subsidized by their parents throughout grad school and afterward.  Morgan et al. (2021) showed that among academics, those with more educated parents received more support and encouragement. This translates to significant wealth gaps between first-gen grads and their peers. A Pew study from last year showed that even among college graduates, first-gen households had lower income (-27%) and much less wealth (-38%) than those whose parents were also college grads. Much of this effect is driven by the fact that kids with wealthy parents incur less or no debt for their education, setting them up for positive wealth–instead of negative wealth–going into graduate school and beyond. In addition to debt avoidance, wealthy parents also confer direct cash subsidies like down-payments for houses and inheritance. All of this means that folks with wealthy parents can accept lower wages.

And it turns out that folks who don’t really have to worry about the monetary benefits of a job are more likely to gravitate to jobs with lower than expected pay, but greater non-monetary benefits like prestige or job security.

Academia is the perfect trifecta of high prestige, high security (with tenure) and low pay. In fact, post-secondary education ranked first in Boar and Lashkari’s (2022) assessment of career intrinsic quality. Folks how receive less than about $50K from their parents (i.e. about the average most academics’ parents give them for a down-payment on a house or to pay for college) are more likely to chose jobs with negative intrinsic quality.

This is compounded by another reason that poor kids are so uncommon in the academy: your parent’s income is the single largest predictor of your early college attainment, far above any other demographic variable (Chetty et al., 2018).

The fact that most folks with PhDs come from wealthy parents with graduate degrees creates a vicious feedback cycle that drives down salaries from graduate student stipends through faculty salaries.

Those low doctoral salaries establish an abysmally low first tier in academic salary ladder. The average salary for a post-doc in the US is $47.5k. (It can be even less appealing internationally. I was offered a European post-doc that would have been < $35K after exchange rate and taxes).

The NSF post-doc salary I turned down would have been $56k, the suggested amount from NSF. For most newly graduated docs, a salary increase to $55k seems enormous to someone who just spent 5 years working 60 hours a week for $30K. But $55K is a paltry salary for someone with a PhD. Consider that the average salary for a professional clown in the US is nearly $50k, and let that subtle irony sink in.

The story outside of academia is much different. Doctorate degrees are actually worth something. And you don’t even have to sell your soul to industry. I got offers for conservation NGO positions for twice my post-doc salary and I interviewed with environmental funding organizations hiring at salaries three to five times my post-doc. I had a career in non-profit leadership before graduate school, so I was probably seeing the high end that the environmental NGO field had to offer. But conservation work is low-paying in general. Other fields have a much higher ceiling.

Right now, the field of data science is booming. Given that even the most field-oriented biologist likely spends most of their days staring at an R terminal doing statistics, every biologist is an experienced and competitive data scientist, de facto. Even in the data science field, you don’t have trade in your morals for your salary. NGOs are also hiring PhDs for data science roles. In fact, the position I ultimately abandoned academia for is with an education non-profit.

I think a lot of academics really want to help make the world a better place through their research. But, the fact that you can make double or triple the salary while doing far more immediately impactful good leaves almost nothing left on the scale in favor of academia.

In my role: I work remotely, have outstanding work-life balance, and a clear promotional track. Compare that to a post-doc where I’d be trying to wrap up unfinished papers from my doctorate on top of a heavy workload, in a temporary position, where the next career step involves competing with over 300 of my peers for a position half-way across the country that pays only slightly more than the people paid to watch Netflix all day.

One of the few benefits that academia can uniquely offer is the promise of tenure. Setting aside the fact that chasing tenure is simply prolonging one’s time chasing a carrot on a stick, tenured positions are dwindling every year. And the chance of transitioning from a post-doc into a tenure track position is abysmal and getting worse. Only about 10% of post-docs end up in tenured roles.

One overlooked downside of academia is that in addition to poor pay, the pay-off is delayed. Sure, in those rare cases you might end up with a six-figure, tenured professorship, but that reward is deferred well past the most important years of capital growth.

For instance, doctoral students and postdoc salaries make it basically impossible to save any money (especially for folks with student loans). As an example, I was making $40k in a full time NGO job right out of undergrad. From that income band, getting paid $33k while getting a PhD seemed like a good bet, given the eventual salary advantage of a degree. However, that calculus neglects to consider the life-timing of savings. The advantage of an extra $7k per year in your early 20s has disproportionate outcomes compared to the same amount later in life.

Consider this, the maximum allowable annual contribution to a Roth Retirement Account is $6k. If the paltry pay of grad stipends and postdocs prevents you from contributing to a Roth IRA for 7 years after undergrad, that mere $42k in lost pay compounds to a loss of nearly half a million dollars by retirement ($484k). (Do the calculations yourself: NerdWallet calculator)

And that’s just the cost of deferred retirement savings. The other engine of wealth accumulation in the US is home equity. Did you know: graduate stipends are not considered eligible income when applying for home loans. And don’t expect to wait until a post-doc, either. You need to be in a position for at least two years for it to be eligible income.*

(*There is a caveat that if you have a signed agreement of continued stipend for at least two years, your stipend could be eligible. But that means that you’d have to buy a house at the beginning of your PhD, when most folks can’t afford a down payment (unless you have wealthy parents to float you down-payment. Or unless you have parents who will gift you $300k in cash to buy your house, like one of my colleagues at Yale).

Ultimately, I’m not saying that a career in academia is a poor life-decision. I’m just saying that a career in academia is a poor life decision if you’re from a poor family.

Folks with external subsidies (even minorly) have the liberty to make decision to follow the dream of academia in a way that those of us who have to generate our own income cannot. Academia runs on those external subsidies. If your family can’t float you during your unpaid summer internships, or loan you cash to pay for the conference that you may or may not get reimbursed for 8 months later, or cover the down payment on your eventual house, … etc., you are going to end up way behind in life.

Unfortunately, this reality remains unseen by those currently in the academy. Academics love to worry about inequality, but because most of them are from upper-middle class to rich families, they manage to overlook the enormous impact of wealth inequality in academia. (Hell, most academic institutions actively avoid even collecting the data that would illuminate this reality).

In the end, overlooking the consequence of wealth subsidies leads mentors to encourage any student who shows an interest to pursue academic careers because they confuse what they wish to be true: “The academy is open to all” with what is, in fact, true: “Academia is a terribly unwise career for folks from poor families.”

I contend that mentors should feel ethically compelled to lay out the Sisyphean asymmetry to mentees from low-income backgrounds. If that makes you uncomfortable, the answer is to fix the system, not to mislead mentees with unexamined ‘luxury beliefs.’

Until that system gets fixed, more of us trailer home alumni will keep unhappily walking away from our dream.

[Updated 2022 Sept. 09 with some recent, relevant research papers]

[Updated 2022 Dec. 10 with information from this Pew report]

 References:

Boar, C., and Lashkari, D. (2022). Occupational Choice and the Intergenerational Mobility of Welfare. Available at: https://www.nber.org/system/files/working_papers/w29381/w29381.pdf.

Cheng, S. D. (2021). What’s Another Year? The Lengthening Training and Career Paths of Scientists. in (Harvard University Department of Economics). Available at: https://conference.nber.org/conf_papers/f159298.pdf.

Chetty, R., Hendren, N., Jones, M. R., and Porter, S. R. (2018). Race and Economic Opportunity in the United States: An Intergenerational Perspective. doi: 10.3386/w24441.

Morgan, A., Clauset, A., Larremore, D., LaBerge, N., and Galesic, M. (2021). Socioeconomic Roots of Academic Faculty. doi: 10.31235/osf.io/6wjxc.

Schultz, R., Stansbury, A., Albright, A., Bleemer, Z., Cheng, S., Fernández, R., et al. (2022). 22-4 socioeconomic diversity of economics PhDs. Available at: https://www.piie.com/sites/default/files/documents/wp22-4.pdf.

The other day, I wanted to tweet out a map showing the distribution of some wood frog tissue samples compared to the entire range of the species. I’m not much for GIS and I didn’t need anything complicated so I wanted to plot in R. If you find yourself in a similar situation needing to plot a simple range map with GGplot, then this post is for you.

This post WILL NOT explain how to estimate a species range. There’s a lot more involved in that process and there are dedicated packages like rangemappr to help.

In this post, we will take a pre-made polygon of a species range and slap it onto a basemap.

The first task is to find a polygon of the range of your favorite species. Unless you already have a shapefile in your possession, one of the easiest places to acquire one is through the IUCN database. You can follow the links to go down the taxonomic rabbit hole for specific species. For this tutorial I’ll be using the red salamander (Pseudotriton ruber), just because I think that they are gorgeous.

Here’s the  IUCN page for the red salamander. In the top right of the page, you’ll see a drop down button labelled, “Download”. You will want to select and download the “Range data – Polygons (SHP)” option. You will need to agree to some conditions in order to download the file, but it is quick and painless. Once you download the zipped folder, extract it into your project directory.

Here are the packages we will be using.

library(tidyverse)
library(rgdal)
library(ggspatial)

First, we need to do a bit of work to get the shapefiles into R. First, we will use rgdal package to read in the shapefiles. Then, we will coerce that object into a dataframe that can play nicely with GGplot.

# Read in the shapefile
PSRU_shape <- readOGR(dsn = "./PSRU_range", layer = "data_0")
# Coerce into a dataframe to play nicely with GGplot
PSRU_shape_df <- fortify(PSRU_shape)

Now we can start mapping. We will use the ggmap package in GGplot that all comes bundled in the tidyverse. First, we define regions for our basemap. In this case, red salamanders range is entirely within the lower 48 United States. Then we plot the basemap and layer on the range map on top. It’s really that simple.

Low48_map <- map_data("state") # Create basemap form GGplot

ggplot(Low48_map, aes(x = long, y = lat, group = group)) +
geom_polygon() +
geom_polygon(data = PSRU_shape_df, fill = "orangered")

But we can make this look a lot nicer. First off, we can ditch the ugly GGplot default theme and add our own color scheme. We can also force GGplot to use a polyconic projection (you’ll need to make sure that you also have the mapproj package installed).

ggplot(Low48_map, aes(x = long, y = lat, group = group)) +
geom_polygon(fill = "cornsilk4", col = "cornsilk") +
geom_polygon(data = PSRU_shape_df, fill = "darkorange3", alpha = 0.8) +
theme_void() +
theme(panel.background = element_rect(fill = "cornsilk")) +
coord_map(projection = "polyconic")

Beyond the basics:

For species with larger ranges, you might want to incorporate more of the continent. We can easily do that by defining larger regions for our basemap. Here, I am still plotting the Lower 48 states basemap to get the state borders.

NAm_map <- map_data("world", region = c("Mexico", "Canada")) # Create basemap form GGplot

ggplot(NAm_map, aes(x = long, y = lat, group = group)) +
geom_polygon(fill = "cornsilk4", col = "cornsilk") +
geom_polygon(data = Low48_map, fill = "cornsilk4", col = "cornsilk") +
geom_polygon(data = PSRU_shape_df, fill = "darkorange3", alpha = 0.8) +
theme_void() +
theme(panel.background = element_rect(fill = "cornsilk")) +
coord_map(projection = "gilbert")

One issue is that it is extremely difficult to crop in if, for instance, we don’t need all of Canada’s norther islands. Using xlim and ylim truncates the polygons in weird ways. And you should NEVER crop with xlim/ylim, anyway.

ggplot(NAm_map, aes(x = long, y = lat, group = group)) +
geom_polygon(fill = "cornsilk4", col = "cornsilk") +
geom_polygon(data = Low48_map, fill = "cornsilk4", col = "cornsilk") +
geom_polygon(data = PSRU_shape_df, fill = "darkorange3", alpha = 0.8) +
theme_void() +
theme(panel.background = element_rect(fill = "cornsilk")) +
coord_map(projection = "gilbert") +
xlim(-125, -60) +
ylim(25, 64)

The better way to crop is to set the xlim and ylim of the clipping mask with coord_cartesian(). But, then we lose the projection.

ggplot(NAm_map, aes(x = long, y = lat, group = group)) +
geom_polygon(fill = "cornsilk4", col = "cornsilk") +
geom_polygon(data = Low48_map, fill = "cornsilk4", col = "cornsilk") +
geom_polygon(data = PSRU_shape_df, fill = "darkorange3", alpha = 0.8) +
theme_void() +
theme(panel.background = element_rect(fill = "cornsilk")) +
coord_map(projection = "mercator") +
coord_cartesian(xlim = c(-125, -60), ylim = c(25, 52))

There are a handful of work-arounds to the problem.

First, there is a code workaround here, but it is a bit complicated.

Second, you can simply export the image at an appropriate size to get a resulting ratio that works. One potential issue with this is that GGplot exports ALL of the information in the PDF, even the polygons outside of the image. In this case, that means a lot of data points to outlines all those tiny Arctic Islands. This can make for some really large PDF files. A very slick solution to this is to rasterize the maps with ggraster before exporting the images. You can still export the image at twice or three times the size to retain high resolution, but the file size will be much smaller.

Third, you could export the uncropped version as a pdf (which is a vector graphic) and open it in a vector graphics program like Adobe Illustrator or Inkscape. From there, you can crop in where ever you want.

Take me to the photos!

Step 1: Come up with a hair-brained scheme.

My labmate Yara and I had been dreaming up the idea studying wood frog genomes from across the species’ range since she started her PhD. Wood frogs have the largest range of any North American amphibian. They also happen to be the only North American amphibian that can survive North of the Arctic circle.

Dr. Julie Lee-Yaw had done a similar study back in 2008. She embarked on a road trip from Quebec all the way up to Alaska to collect wood frog tissue. So, out first step was to ask Dr. Lee-Yaw if she would collaborate and share her samples.

Those samples gave us a solid backbone across the wood frog range, but we were missing population in expansive regions north and west of the road systems. We worked with the Peabody Museum to search for tissue samples that were already housed in natural history collections around the world. We filled a few gaps, but huge portions of the range were still missing.

We knew that there must be samples out there sitting in freezers and labrooms that were not catalogued in museum databases. So, our next step was to begin sleuthing. We looked up author lists from papers and cold-called leads. I even reached out to friends on Facebook (…which actually turned out to be a big success. The aunt of a friend from undergrad happens to do herpetology research in Galena, Alaska and was able to collect fresh samples for us this year!). This effort greatly expanded our sample coverage with new connections (and friends) from Inuvik and Norman Wells in the Northwest Territories, Churchill on the Hudson Bay, and the Stikine River Delta in Southeast Alaska.

But as the points accumulated on the map, we noticed some glaring holes in our coverage. Most importantly, we had no samples from Northwestern Alaska. Populations in this region are the most distant from the ancestral origin of all wood frogs in the southern Great Lakes. If we wanted a truly “range-wide” representation of wood frog samples, we needed tissue from that blank spot on the map!

Step 2: Convince your advisor and funders it’s a good idea.

This might be the hardest step. In our case, Yara and I were lucky that our advisor, Dave, was immediately supportive of the project. After we made the case for the importance of these samples, funders came around to the idea as well.

Step 3: Make a plan …then remake it …then make a new plan yet again.

Once we knew where we required samples from, we needed to figure out how to get there. Alaska in general is remote, but northwestern Alaska is REALLY remote. The road system doesn’t stretch farther than the middle of the state. All of the communities–mainly small villages–are only accessible by plane, and most of them only have runways for tiny prop planes. Travelling out from the villages into the bush is another layer of difficulty. Most people here either travel by boat on the river or by snowmachine during the winter. Traveling on land, over the soggy and brush-choked permafrost, is brutal and most locals only do it when necessary, if at all.

Prior to academia, I made a career of organizing expeditions to the most remote places in the rugged southeastern archipelago of Alaska. Despite my background, the logistic in the Arctic were even inscrutable to me. Fortunately, I had a couple of friends, Nick Jans and Seth Kantner, who know the area well. In fact, Seth grew up in a cabin out on the Kobuk. (Seth and Nick are both talented authors. I suggest checking out Ordinary Wolves by Seth and The Last Light Breaking by Nick). With their help, I was able to piece together the skeleton of a trip.

After many logistic iterations, Yara and I decided to follow in the footsteps of local hunters who, for generations, have used the rivers as conduits into the heart of the wilderness. Our plan was to travel down one of the major arterial rivers and hike inland to search for frog as we went.

Our original itinerary was to raft the 100 mile section of the Kobuk River from just north of Ambler village to the village of Kiana. But at the last minute (literally), our plans changed. As we were loading up the plane, the pilot told us that he couldn’t fly into our planned starting point. Instead, he suggested that we fly into a gravel bar 30 miles up river in Gate of the Arctic. Those “30 miles” turn out to be AIR MILES. Following the river, it ended up adding over 60 miles to our trip.

We packed two inflatable oar rafts, almost 150 pounds of food, and another 300 pounds of camping, rescue, and science gear, into the balloon-wheeled plane. For the next two weeks, we rowed down the swift Ambler River from the headwaters to the confluence of the Kobuk. Then, we rowed down the massively wide and meandering Kobuk River, eventually extending our trip by an additional 30 miles, by-passing Kiana, and continuing to Noorvik, the last village on the river.

Step 4: Recruit a crew.

Despite being the worlds first and only Saudi Arabian Arctic Ecologist with limited camping experience, I knew Yara would be a stellar field partner. But I never like traveling in brown bear country with fewer than four people. Plus, expedition research involves too many daily chores for the two of us to manage alone. So, we recruited a team.

Sam Jordan is a dry land ecologist, but he had been willing to help me with my dissertation fieldwork in wetlands before, so I knew he would be willing to defect for a good adventure. Sam is also an exceptional whitewater paddler and all-around outdoor guru. Plus, he’s just a great guy (when he leaves his banjo at home). He and I spend two weeks floating the Grand Canyon in the dead of winter and there are few people I would want along on a remote river trip.

Kaylyn Messer and I guided sea kayak expeditions in Southeast Alaska back in our youth. I am a bit particular about how I manage my camp system (read: “extremely picky and fastidious to a fault”) on big trips. Kaylyn is one of the few people as scrupulous as me, but she’s also a super amenable Midwesterner at heart. I knew she’d be a huge help out in the field.

We fell into an effective rhythm on the trip.  Each morning we woke, made breakfast, broke camp, packed the boats, and launched early in the day. While one person on each boat rowed, the other person checked the maps for frog surveying spots, fished, or photographed. We stopped along the way to bushwhack back into wetlands we’d identified from satellite images. We typically arrived at camp late. Yara and I would set up one tent to process the specimens from the day while Same and Kay made camp and cooked dinner. One of the hidden disadvantages of 24-hour Arctic sunlight is that it is easy to overwork. Most nights we only managed to get sampled finished, dinner cleaned up, and camp bearproofed with enough time to crawl into tents with just eight hours till beginning again the next day.

Step 5: Do the science.

Doing science in the field is difficult. Tedious dissections seem impossible while baking in the omnipresent sun and being alternately hounded by hundreds of mosquitoes or blasted by windblown sand. Trading lab coats for rain jackets and benchtops for sleeping pads covered in trashbags compounds the trouble. Not to mention, keeping tissues safe and cool. Organization and adaptability go a long way.

On remote, self-supported trips, it is inevitable that equipment fails or is lost. On one of the first days, we discovered that our formalin jar was leaking—and formalin is not something you want sloshing around! We cleaned the boats and found a creative solution to replace the offending container: a 750ml Jack Daniel’s bottle!

Planning ahead and engineering backup plans also helps. One of our main struggles was figuring out how to preserve specimens and get them home. It is illegal to ship alcohol by mail and you can’t fly with the high-proof alcohol needed for genetic samples. You can ship formalin, but it is difficult to fly with. To make matters worse, we were flying in and out of “dry” or “damp” villages where alcohol is strictly regulated or forbidden. Also, we happened to be flying out on a Sunday, making it impossible to mail samples home. The solution we arrived at was to ship RNAlater and formaldehyde to our hotel room ahead of time. Tissue would remain stable in RNAlater for a couple of weeks and we could make formalin to fix the specimens. After fixing, we cycled the specimens through water to leach out the formalin. This made it possible for me to fly with all of the tissue tubes and damp specimens in my carry on. Other than a few concerned looks from the TSA folks, all of the samples made it back without issue!

Step 6: Enjoy the adventure.

Despite the hard work, there was a lot to appreciate about the Arctic. We witnessed major changes in ecology as we travelled from the steep headwater streams in the mountains to the gigantic Kobuk. Every day was an entirely new scene.

Here’s an illustrated thread on our collecting trip to the Alaskan Arctic for the @Yale @yalepeabody this past summer. 1/22 pic.twitter.com/Jh3VCqVEgx

— A. Z. Andis Arietta (@azandisarietta) August 24, 2021

Step 7: Forget the hardships

Looking back, it is really easy to forget the sweltering heat, swarms of mosquitoes, inescapable sun, and freak lightning storms. And, it’s probably better to forget those anyway!

Arctic CircleSam illustrates that we are, in fact, above the Arctic Circle.

GearGear (minus food and rafts) ready for weigh-in before loading the plane.

Upper Ambler RiverOur drop-off point on the braided upper reach of the Ambler River in Gate of the Arctic National Park.

Gates of the ArcticGetting dropped off on a gravel bar in Gates of the Arctic National Park.

The first dayUnloading the plane on the first day of the trip.

Camp, night 1Campsite along the Kobuk River.

Rigging the raftsPacking up the two-person rafts on the first day.

Kaylyn Messer at the helm.In the upper rivers, near the headwaters, the current was pleasantly swift.

Upper riverSam and Yara rowing down the Kobuk River.

Early morning rowingRowing the swiftwater on the headwaters of the Ambler River in Gates of the Arctic National Park.

Searching for frogs

First frogs!Yara and I after finding the first frogs of the trip!

Science in the fieldYara preparing to do the sciencing with a garbage back as a makeshift lab bench.

Dinner. night 6Dinner and a view on the banks of the Kobuk River.

Midnight sunTrying to avoid getting a sunburn while sleeping in the Arctic.

Midnight sunThe sun baking our tent at 1130pm.

Rainy rowingSam and Yara floating downstream.

Arctic frogYara proudly presents her first wood frog collection.

Kobuk viewsSmoke and clouds mute the midnight sun.

Sam JordanSam Jordan in his happy place.

Chocolate con culicidae.With hundreds of bugs circling your head at all times, it is impossible not to end up with a few in every meal.

CutbanksMassive cutbanks made much of the shore inaccessible for sampling.

StormA freak lightning storm.

SunsetBeautiful light from the midnight sun.

Field dissection Taking tissue samples for sequencing.

TadpolesTadpoles from "Seth's Pond".

Prepping collectionsYara and I taking tissue samples and prepping specimens.

Arctic Wood FrogBeautiful arctic wood frog.

Fixing specimensHere, I am fixing specimens with formalin to preserve them for the Peabody Museum.

Bushwhacking The view while bushwhacking in to search for frogs.

BugsA random sample of friendly mosquitoes.

BearGrizzly bear prints on the beach.

FishingSam with his first of many greyling.

The Kobuk River valleyThe Kobuk Valley watershed is just an endless expanse of wetlands.

The end of the tripFlying back to Kotzebue in a packed plane at the end of the trip.

Heading homeCatching a ride out to the airstrip with our truckload of gear on the last day of the trip.

I’ve always been enamored with books, but when I started my PhD, I was worried that I would fall out of the habit of reading for fun. So, I set a goal for myself: read all of the Pulitzer Prize winning fiction novels before I finished my degree.

The Prize for Novel was one of the original Pulitzers. The first was awarded in 1918 and the competition has run annually since, although the category name was changed from Novel to Fiction in 1947. In eight years, no prize was awarded (the last time was in 1977 when the committee passed up Norman MacLean’s A River Runs Through It).

So, in total, there are 93 winning novels as of 2020. A few (8), I had read prior to setting my personal challenge. In the end, it took me 5 years to read all 36,518 pages. Some of my favorites were Ironweed, A Confederacy of Dunces, A Bell for Adano, Middlesex, Arrowsmith, All the Light We Cannot See, So Big, and Laughing Boy. I’ve included my rating for all of the novels below.

I really enjoyed making reading-for-fun an objective challenge. It certainly coerced me to read more. Surprisingly, this challenge not only helped me maintain my reading habit but increased my consumption. Between each Pulitzer, I generally read either a non-fiction book or sci-fi/fantasy novel, which I also track. In the end, I managed to read an average of slightly over 3 books a month. Given the constant reading required for my degree, there is no way I would have read a fraction of those books without adding time for reading to my to-do list for the week.

With this challenge over, I am ready to start another. To be honest, I was a bit disappointed in the Pulitzers for including a fair portion of really boring books (I will NEVER read another John Updike novel in my life). So, for my next challenge, I’ve decided to aggregate rankings from as many “Books to Read Before You Die” lists as possible and read the top 100. I’ll publish a follow up post once I generate that list.

Ratings

Books are arranged by award year within ranks. Here’s the thought process behind my ratings:

5 = I would re-read this book

4 = I would recommend this book

3 = I’m glad I read this book

2 = I would recommend against reading this book

1 = I regret the time I spent reading this book

5

So Big – Ferber (1925)

Arrowsmith – Lewis (1926)

Laughing Boy – La Farge (1930)

The Good Earth – Buck (1932)

The Grapes of Wrath – Steinbeck (1940)

A Bell for Adano – Hersey (1945)

The Old Man and the Sea – Hemingway (1953)

To Kill a Mocking Bird – Lee (1961)

Angle of Repose – Stegner (1972)

A Confederacy of Dunces – Toole (1981)

The Color Purple – Walker  (1983)

Ironweed – Kennedy (1984)

A Good Scent from a Strange Mountain – Bulter (1993)

The Amazing Adventure of Kavalier and Clay – Chabon (2001)

Middlesex – Eugenides (2003)

The Road – McCarthy (2007)

All the Light We Cannot See – Doerr (2015)

Empire Falls – Russo (2002)

4

The Bridge of San Luis Rey – Wilder (1928)

Gone with the Wind – Mitchell (1937)

The Yearling – Rawlings (1939)

Dragon’s Teeth – Sinclair (1943)

Tales of the South Pacific – Michener (1948)

The Way West – Gutherie (1950)

The Caine Mutiny – Wouk (1952)

A Death in the Family – Agee (1958)

Advise and Consent – Drury (1960)

The Edge of Sadness – O’Connor (1962)

The Reivers – Faulkner (1963)

The Fixer – Malamud (1967)

The Stories of John Cheever – Cheever (1979)

Foreign Affairs – Lurie (1985)

The Shipping News – Proulx (1994)

Martin Dressler: The Tale of an American Dreamer – Millhauser (1997)

The Brief Wonderous Life of Oscar Wao – Diaz (2008)

The Overstory – Powers (2019)

The Nickel Boys – Whitehead (2020)

His Family – Poole (1918)

All the King’s Men – Warren (1947)

The Town – Richter (1951)

The Collected Stories of Katherine Anne Porter – Porter (1966)

The Killer Angels – Shaara (1975)

A Thousand Acres – Smiley (1992)

A Visit from the Goon Squad – Egan (2011)

The Goldfinch – Tartt (2014)

Underground Railroad  – Whithead (2017)

3

Alice Adams – Tarkington (1922)

The Able McLaughlins – Wilson (1924)

Early Autumn – Bromfield (1927)

Scarlet Sister Mary – Peterkin (1929)

Years of Grace – Barnes (1931)

The Store – Stribling (1933)

Lamb in His Bosom – Miller (1934)

The Late George Apley – Marquand (1938)

Journey in the Dark – Flavin (1944)

Andersonville – Kantor (1956)

The Travels of Jaimie McPheeterrs – Taylor (1959)

The Keepers of the House – Grau (1965)

The Confession of Nat Turner – Styron (1968)

The Collected Stories of Jean Stafford – Stafford (1970)

Elbow Room – McPherson (1978)

The Executioner’s Song – Mailer (1980)

Lonesome Dove – McMurty (1986)

The Hours – Cunningham (1999)

The Known World – Jones (2004)

Gilead – Robinson (2005)

March – Brooks (2006)

Tinkers – Harding (2010)

Less – Greer (2018)

Now in November – Johnson (1935)

Beloved – Morrison (1988)

Breathing Lessons – Tyler (1989)

Independence Day – Ford (1996)

The Orphan Master’s Son – Johnson (2013)

2

The Magnificent Ambersons – Tarkington (1919)

The Age of Innocence – Wharton (1921)

One of Ours – Cather (1923)

Honey in the Horn – Davis (1936)

In this Our Life – Glasgow (1942)

Guards of Honor – Cozzens (1949)

A Fable – Faulkner (1955)

The Optimist’s Daughter – Welty (1973)

The Mambo Kings Play Songs of Love – Hijuelos (1990)

Rabbit at Rest – Updike (1991)

The Stones Diaries – Shields (1995)

American Pastoral – Roth (1998)

The Sympathizers – Nguyen (2016)

1

House Made of Dawn – Momaday (1969)

Humboldt’s Gift – Bellow (1976)

Rabbit is Rich – Updike (1982)

A Summons to Memphis – Taylor (1987)

Interpreter of Maladies – Lahiri (2000)

Olive Kitteridge – Strout (2009)

Today it is trivial to make really sexy graphics in a few clicks and keystrokes. But the ease of creation also makes it much easier to produce poorly planned or spurious outputs. It also means that the marketplace of people’s attention is now flooded with loads of other eye-catching data visualizations to compete with.

Now, more than ever, it is important to think strategically about how to present your work. This blog grew out of a guest lecture I gave. It is intended to present some conceptual tools to help you make your data stand out.

To data viz or not to data viz?

Making a stellar data visualization takes time and effort. Even a simple plot for a scientific paper can take a while to get to the final print-ready stage. So, for starters, it is worth considering just how much time your particular data viz project is worth.

It is really easy to go deep into a rabbit hole making a beautiful visualization, or even an entire data storytelling project, only to have it sit on your computer or collect digital dust in some dark corner of your blog. My most adamant piece of communications advice is that you should spend just as much time planning how to outreach your work as you do creating it. And after you’ve got your product, you should again spend the same amount of time actually making sure that people see it. (This rule applies less to journal figures since the outlet is predetermined, but you should still plan to spend as much time sharing your hot-off-the press manuscript with your stunning figures after it comes out.)

Data visualization or data storytelling?

When people think about great data visualizations, they often think about the flashy and interactive products like those from the Washington Post or New York Times. I also love these interactive visuals, but to me, they are something more than data viz–they are data storytelling. Rather than simply displaying data, data storytelling integrates data as a part of a larger narrative. Good data storytelling involves skills that overlap with data viz, but add much more. For instance, my friend Collin’s Story Map of his research on lizards evolving to hurricanes is a great example. We learn all about his research and how he produced his data, but very little about the data itself.

In this article, I want to focus narrowly on data viz and how we interpret statistics visually. There are loads of plot forms that you can use, and folks are always coming up with new ways to use them, so rather than create an exhaustive list, I want to consider when and how we use data visualizations.

 One quick caveat here: data viz implies that the only way to interpret data is with sight. But there are some really cool projects that display data without visuals, like my friend Lauren, who translated Alaskan tree loss through sound.

Grabbing your attention or focusing your attention?

One of the first questions to ask yourself in defining the purpose of your visual is: am I trying to grab folks’ attention or do I want to focus their attention? Humans brains are not all that well designed for sustained attention (I go in depth about this in my presentation about scientific presentations), so most of our task as science communicators is simply managing people’s attention spans. Flashy and interactive visuals are great for catching your audience’s eyes, but can be a distraction from carefully interrogating specific trends in data because there is too much to focus on. On the flip side, an equally beautiful but more subdued plot can perfectly highlight a specific point you want to make about your data, but folks might flip or scroll right past it if they are not actively interested. Considering your audience is paramount. For example, in a paper, I may include lots of information in a plot, but when I present my work in presentation form at conferences, I completely strip down my figure to their most basic elements.

One of the reasons we have short attention spans is that our brains have evolved to process lots of information quickly. As a tradeoff, our brains take cognitive shortcuts. If we are clever, we can use visualizations to hack our brains and leverage those shortcuts. As an example, take a look at the two images below. Can you tell which image of stars is randomly placed? Can you tell which set of numbers is random?

Human brains are overly tuned to seek patterns. Often, we see patterns when none are there (maybe this is where human predilection for superstition, conspiracy theories, and religion come from). Most people think that the blue stars (B) and number string A are the random sets. That is because we tend to see too much pattern and clustering in the black stars and too many patterns of repeats in number string B. When we see patterns, we assign meaning. In fact, the black stars are randomly placed (the blue stars are overly uniform) and number string A is randomly generated.

This is convenient for data viz, because it makes it easy for us to see trends in complicated data. For example, when Nature plotted all of it’s published papers over the last 150 years, and then linked them by citations, the result was incredibly complicated. But our minds tune-out most of the noise and instead focuses in on the major groups where fields merge. 

On the flip-side, our minds are quick to spot deviations from patterns, too. For instance, when Campbell et al. plot coding density versus genome size, it is easy to spot the clade of endosymbionts (in green) that deviate from the trend.

Our brains are also really bad at conceptualizing large numbers. For instance, if I told you that humans have about 3.2 billion bits of information in every cell of your body, but E. coli has just 5 million, and Paris japonica flower has almost 150 billion, the scale might be hard to grasp. But if I compare your genome to the letters in an encyclopedia and visualize the difference, the disparity is clear.

Making visuals that strategically hack our brains.

When it comes to visuals, I don’t like prescribing rules. Aesthetics change too quickly. Instead, I think it is more helpful to be strategic about the content of your visuals and treat the aesthetic refinement as an artistic process. 

Scott Berinato’s book Good Charts comes from the perspective of management rather than science, but is, nonetheless, one of the best examples I’ve found of thinking strategically about making visuals. Berinato thinks that visuals fall on two intersecting gradients: Conceptual versus Data-driven (are you dealing with ideas or statistics?) and Exploratory versus Declarative (are you looking for a pattern or are you showing a pattern?).

1. Everyday data viz

Usually, when we think about data viz, we are thinking about graphics that fall into the upper right quadrant, data-drive declarative graphics, what Berinato calls “Everyday data viz.” The purpose of these graphics is to highlight specific facts about our data. Most of the figures from scientific papers fall into this category.

Within the “everyday data viz” category, there lies a wide range of visualization goals that depends on the intended audience. For example, I made a mirrored radial barplot comparing population density to tree cover. Wrapping this plot into a radial form makes the data more interesting, but actually makes it more difficult to read. If I were to include these data in a scientific paper, I would probably use a dotplot like the one in the top left of the figure. The dotplot displays the same information in a way that is more conducive to quantitative comparison.

With these types of visuals, there is often a tradeoff between simplicity and aesthetics. Usually, simpler is better for scientific audiences. However, sometimes the whole point of a graphic is to demonstrate complexity or variation in the data. For instance, a simple mixed model regression could be easily displayed as a single trend line.

Not only is this super boring, but it misses one of the points of mixed models, which is how we deal with variation in the data. Below are six examples showing the same trend while highlighting the variation in the data in different ways.

On the other hand, when giving scientific presentations, we want to highlight the main trend without distracting the audience with noisy variation. In a prior post, I used the fake example below, where the most important trends (bottom figure) are completely buried in the meaningless distraction of too much information (top figures).

My main point here is that you must be strategic about who your audience is and exactly what you want them to take away from your visuals. It is unlikely that anyone will think as carefully about your graphic as you have. Instead, most folks will take away a fraction of the information you present. So, it is worth being as parsimonious as possible with the content in your graphics. One tip for presentations is to step away from your computer and squint your eyes–if you can’t make out the main trend, you probably should strip it down. Another tip is to start with the bare axis and explain them to your audience before showing the content of the plot. This way, they already know what to expect and they will not be as distracted trying to conceptualize what the graphic is saying.

2. Visual discovery

The graphics in the upper right quadrant of Berinato’s diagram are like the perfected Pintrest versions of our visuals. Before we get to that point, we will probably plot a ton of graphs as we analyze our data that no one ever sees. Berinato calls these graphs “visual discovery.” They fall in the lower right quadrant of data-driven exploratory plots. 

As we explore our raw data, it is useful to hack our own brains to discover hidden patterns in our data. Most data is multidimensional and too complex to see every relationship at once. So, we check for relationships among variables and among subsets of variables. This process is usually iterative. The point isn’t to make perfect, pretty graphics–the point is to wrap our minds around the data.

One of my favorite examples of visual discovery involves one of the oldest examples of data viz. 

In the mid 1800s cholera was sweeping into London. At the time, few understood how the disease was transmitted. John Snow (no, not that John Snow) a medical doctor decided to plot the cases as bar charts of the number of victims at each address on a street map of the city. The map showed a public well at the center of the epidemic. The map helped Snow convince skeptical municipal authorities to close the well and effectively ended the outbreak.

Visual discovery is what scientist probably spend 80% of their analysis time on (I certainly do). Plotting programs like Rstudio or MatLab (and to a lesser extent, Excel) make it really easy to play with lots of ways to see our data and easily iterate to narrow in on interesting trends.

3. Idea illustration

The top left quadrant, conceptual and declarative, Berinato calls “Idea illustration.” These are usually heuristics, flow charts, or diagrams with the purpose of visually demonstrating a complex idea in picture form. Scientists use these type of graphic often in review or synthesis papers. For example, I made the figures below for a recent review paper of herp thermal evolution. Neither are based on data. The first demonstrates a theoretical process. The second illustrates what real data might look like and how to interpret them. These types of graphic hack the map reading tendencies of our brain or prime our natural pattern seeking.

4. Idea generation

The lower left quadrant, Berinato calls, “Idea generation.” These are the kinds of figures scientists scribble up on white boards when we are thinking through experiments. Rarely do these graphics make it out into the world, rather they help us think through our own ideas. However, sometimes conceptual, exploratory graphics are useful for thinking through hypotheses. For example, I included the graphic below in my dissertation prospectus as a way to think through how geneflow patterns might look in different populations.

Understanding why and how it makes sense to use graphics can save you loads of time, keep you from making spurious plots, and may even lead you to a new discovery. Fortunately, professional plotting tools like (R and GIMP2) are freely available. So get out there and start making something beautiful and useful!

I recently wrote a paper looking at how frog breeding timing is impacted by climate change. So, I’ve been reading lots of ecological studies of phenology (more on phenology later). One thing that struck me is how almost everyone in ecology misuses the term “Julian Day” when they mean Day-of-Year.

Day-of-Year (DOY), as the name suggests, is the count number of a given day in the year. So, Jan 25 is DOY 25 and March 1 is either DOY 60 or DOY 61 depending if it is a leap year. And we can express the time of day as a decimal, so that 3pm on January 1 is DOY 1.625.

Julian day is a completely different way to measure time. It was defined by an astronomer named Joseph Scalinger back in 1583 (and so, takes serious precedent over contemporary ecologists trying to hijack the term).

The point is, DOY and Julian day/date are wildly different things designed to measure wildly different phenomena.

Unlike DOY that starts counting on January 1^st in any given year, the Julian day count starts on January 1, 4713 BC. There is a complicated historical reason that Scalinger chose 4713 as the starting date that had to do with wedding the Julian and Gregorian dates during the calendar reform (read all about that here), but the point is, DOY and Julian day/date are wildly different things designed to measure wildly different phenomena.

For instance, I’m writing this blog on the 25^th of January 2020.

The DOY today is: 25

The Julian day today is: 2458873

But, it gets even crazier because unlike the DOY count that starts at midnight, Julian days start counting at Noon. So, right now, at 1030am the Julian day is 2458873, but after lunch it will be 2458874.

The Julian day metric is essentially worthless for comparing seasons. There is no ecologist who uses true Julian days; so, please, ecologist, don’t say Julian Day when you mean Day-of-Year.

As Gernot Winkler, former USNO Timer Service director notes:

“[Mixing Julian Day and DOY] is a grossly misleading practice that was introduced by some who were simply ignorant and too careless to learn the proper terminology. It creates a confusion which should not be taken lightly. Moreover, a continuation of the use of expressions “Julian” or “J” day in the sense of a Gregorian Date will make matters even worse. It will inevitably lead to dangerous mistakes, increased confusion, and it will eventually destroy whatever standard practices exist.”

So why does everyone misuse Julian Day? My hunch is that Julian Day sounds more technical than DOY, so folks gravitate toward it and others follow suit without ever questioning what it means.

Why do we care about studying seasonal change across years?

Phenology is the study of seasonal cycles of lifehistory like when bears go into hibernation, when flowers open, or when geese migrate. Phenology is a hot topic these days because climate change is causing wild populations to change their seasonal timing (Thackeray et al. 2016). For instance, frogs increasingly start calling and breeding earlier (Li et al. 2013) and forests green-up earlier (Cleland et al. 2007).

On one hand, shifts in lifehistory timing might be a good way to cope with climate change, but it can be bad news if shifts in one species causes a misalignment in an ecological relationship (Miller-Rushing et al. 2010; Visser & Gienapp 2019). For example, European flycatcher migration generally coincides with a boom in caterpillars that feed on oaks. However, climate change drives oaks to bud earlier, which means that all the juicy caterpillars turn chrysalises before the birds show up (Both & Visser 2001; Both et al. 2006). Similarly, snowshoe hares evolved to change coat color from white to brown in winter, but as snow melts earlier and earlier each year, rabbits are stuck with white coats for too long and become easy targets for predators (Mills et al. 2018).

Needless to say, it is important for use to be able to compare when in the season these critical phenomena take place and compare their change across years. When we do so, we are using DOY to align datasets across year, not Julian day; so, ecologists, let’s stop using the wrong term.

References:

Both, C., Bouwhuis, S., Lessells, C. M., and Visser, M. E. (2006). Climate change and population declines in a long-distance migratory bird. Nature 441, 81–83. 

Both, C., and Visser, M. E. (2001). Adjustment to climate change is constrained by arrival date in a long-distance migrant bird. Nature 411, 296–298. 

Cleland, E. E., Chuine, I., Menzel, A., Mooney, H. A., and Schwartz, M. D. (2007). Shifting plant phenology in response to global change. Trends Ecol. Evol. 22, 357–365. 

Li, Y., Cohen, J. M., and Rohr, J. R. (2013). Review and synthesis of the effects of climate change on amphibians. Integr. Zool. 8, 145–161. 

Miller-Rushing, A. J., Høye, T. T., Inouye, D. W., and Post, E. (2010). The effects of phenological mismatches on demography. Philos. Trans. R. Soc. Lond. B Biol. Sci. 365, 3177–3186. 

Mills, L. S., Bragina, E. V., Kumar, A. V., Zimova, M., Lafferty, D. J. R., Feltner, J., et al. (2018). Winter color polymorphisms identify global hot spots for evolutionary rescue from climate change. Science 359, 1033–1036. 

Thackeray, S. J., Henrys, P. A., Hemming, D., Bell, J. R., Botham, M. S., Burthe, S., et al. (2016). Phenological sensitivity to climate across taxa and trophic levels. Nature 535, 241–245. 

Visser, M. E., and Gienapp, P. (2019). Evolutionary and demographic consequences of phenological mismatches. Nat Ecol Evol 3, 879–885. 

The featured image of this post is from joiseyshowaa under creative commons usage.

I’ve replicated all of the slides and script below, but first, here are the bullets if you’re short on time.

To make better presentations, try:

• Reorient your presentation to optimize for short attention spans. This means stating your conclusion first, then presenting your data, then connecting it to the broader topic.
• Minimize distractions. Avoid designs and presentation strategies that will distract from your main topic.
• Bad storytelling can be a distraction from even the best story, so please, practice your #$*&ing presentation!
• Use figures to tell stories, not to show data. Reduce data stories to minimum story units and only tell one story unit at a time using layers, starting with the axes.
• I also include a number of simple PowerPoint hacks that will help you avoid common errors that lead to wonky slides.

Here’s my full presentation:

What is the point of academic presentations? The obvious answer is “to transmit information about our science.” Realistically, most of us are also motivated by little more than adding a new line to our CV. An underappreciated purpose of presentations is as a vehicle for networking. Presentations are like calling cards and signal to your audience if your work is interesting. It also give audience members a brief window into you as a scientist.

We need to give good presentations because, first, the whole damn point is to transmit our information. Bad presentation are a barrier to that transmission. Second, we need to give good presentations because presentations are like our tinder profiles to collaborators. A bad presentation indicates that you do not care enough about your work or your peers in the audience to bother making a tolerable presentation.

Some might question the wisdom of signing up to give a presentation about not giving bad presentations. But you should be able to guess from the title of my talk that I think the bar is exceptionally low. More often than not, academics give pretty terrible presentations. But there are a few simple things you can do to give really great presentations.

It basically boils down to just two key elements:

Optimize for attention

Minimize distractions

Let me show you what I mean by Optimizing for Attention.

Almost every science presentation I have ever seen is organized the same way that we organize papers.

We start with some big picture background questions, then we talk about our specific question, then we explain our system, then we detail our methods, then we talk through our results, and if we have enough time at the end, we finally get to our conclusion, which is the main punchline of the talk. That is the chocolate at the center of the tootsie-pop.

The problem with this model is that our attention span diminishes over time. So, we end up wasting all of the optimal attention window laying out the least important information, and we wait until half the audience has zoned out before delivering the punchline.

Lindquist & McLean (2011) showed how attention diminishes over time by surveying folks during 45-minute lectures. They sounded a signal at intervals during presentations and then had respondents indicate if they were thinking about something unrelated to the lecture, called Thoughts or Images Unrelated to Tasks (= TUITs (basically, day dreaming)). The frequency of daydreaming increased over time until only about half of the audience was focused at any given time towards the end of the presentation.

To optimize for attention spans, one option is to flip your presentation to match the audience’s daydreaming frequency. Give the punchline first. Rather than starting by explaining why your question matters, start with your conclusion and then explain how it is relevant.

Even though this might seems strange, it is actually how we read papers. Rarely do we ever read papers linearly. Most people skip to the last line of the abstract to get the punchline, then maybe read the abstract or figure captions, then the conclusions. Maybe you read the methods or the text of the results section next, but the last thing you read is the background or introduction.

Even if you arrange your presentation to optimize for folks attention span early, we still don’t want to lose most of our audience along the way. We’d like to keep folks’s attention as much as possible.

We humans are a very distractible animals. We spend 47% of our day distracted by other things. And we are most distracted at the times we are trying to concentrate the most!

This makes maintaining attention even more difficult in learning environments, because we will naturally seek out distraction.

For example, kids in decorated classrooms performed 25 – 35% poorer on tests because of the easy distraction from the environment. Interestingly, even without environmental distraction, kids simply switched to being distracted by each other.

In the case of classrooms, there are likely many other benefits of busy classrooms that outweigh the distractions they cause. But in the short span of a presentation intended for adults, we should be aiming for the starkest, least distracting design.

Just to drive home my point that we humans are overly distractible, look at the next slide and time how long it take you to find the let “O”.

Now try it again with the next column.

When researchers repeat these kinds of tests over and again, they find that just adding a simple distraction, like the cartoon, substantially increases your processing time because your mind splits your attention to processing the distracting image.

So, given what I just told you about how easy it is to distract a human, what is wrong with this fictitious but typical academic style slide?

The problem with this type of slide is that there is too much going on. Even the most aggressively wielded laser pointer will not be able to focus the audience’s attention to one element at a time without distraction.

It is also important to remember that even the best story can be ruined by a bad storyteller. Poorly practiced and desultory presentations can be a huge distraction. So, in addition to minimizing visual distractions, remember to…

In the next section, I will go over some pointers and presentation hacks to avoid slides like this one.

Using figures

Academics love to put up complex figures with loads of distracting and unnecessary information on slides. Which kind of makes sense, really. After all, you spent hours collecting each little point of data, so to you, every data point is important. But that’s often not the case. Our goal with figures is to tell a story. It’s not to show off how much work we did, or how complicated our designs are. We want to distill our figures down to the smallest possible story units.

Take a look at this figure from an “experiment” I conducted. For this fictitious experiment, I was interested in how beard color and length correlates with crossword completion speeds. I went to 16 towns. In each town I gave 100 bearded folks a crossword and recorded how many cells they completed per minute and measure their beard length. In every town, half the folks had red beards and half had brown. I noted if they had proper beards or goatees (type).

Think about the minimum units of the story here. What are some ways you might be able to make this figure simpler and less distracting?

Here is my revision. All of those original cells told basically the same story. And this image answers my primary question: “How does beard length and/or color impact crossword speed?” There is no need for the other information in the figure.

Now, here is how I would present this slide. I do it in layers, starting byexplaining the axis and what we should expect to see.

Then we layer on one bit of information.

Then the next bit of information.

But there are some times when distilling down all of the information into one figure is difficult or doesn’t answer our question.

In this next fictitious experiment, I wanted to know if dragon size correlated with the number of villagers eaten. I recorded three different species of dragons in three different years (years is in time from present).

In this case, there is no obvious story–the relationship changes between years and with different dragon species. Also, our sample size are very different, so we need to have a way of relating that we are more confident in some relationships than others.

Here is how I decided to tell this data story.

First, I start with the blank axis and explain what they mean.

Then I add the first year of information.

Then, I add the second year of information. But I still want folks to see the prior information, so I use selective highlighting to focus their attention. We can see that in all cases, more villagers were consumed. And the rate of growth increased.

Now we add the third element. And we can tell the whole story. Every year, dragons eat more villagers, but species differ in the level of increase. Also, the relationship changes over time with respect to body size for different species of dragon.

To recap the whole story, I might show just the trends lines and confidence bands.

It can be really hard to strip down figures, especially if you are afraid that someone might question your data. When I make presentations, I always keep all of my original figures in extra slides at the end of my presentation, after the conclusion slide. I never show those slides in the presentation, but if someone asks a specific question about the data, I can easily flip to the more informative figure.

Other trick and tips

You can use the same kind of selective color to focus attention with text, too.

I also want to touch on some problems that I see too often and tell you how to avoid them.

Have you ever seen a presentation that looked fine on your computer, but came out looking like this at your conference talk?

This happens when you use a font on your computer to make a presentation that is not installed on the computer that you use to display the presentation. The computer defaults to what it thinks the next closest font should be, and it is always wrong.

The easiest solution is to simply export your slides as JPEG image files and put each one back onto a slide. Essentially, your slide is now a picture of your original slide. That way, it will be displayed exactly as you see it on your computer wherever you display it.

One quick word while we are talking about fonts. Please try to pick simple fonts (like those on the left). You should only use the fonts on the right if you are creating a title page for your 5th grade history report.

Have seen presentations with figures that look like this?

This happens when you enlarge an image in bitmap format. Essentially you are trying to display more pixel than in the original image. The computer interpolates new pixels by averaging adjacent pixels, but it is fuzzy and pixelated. The solution here is to either use vector based images or bitmaps that are as large or larger than the display size. If you are using an image from a paper, try to download the largest image size. If you can only take a screen shot, be sure to blow it up as large as possible before capturing the screen.

And in conclusion…

Please don’t do this at the end of your presentation. No one needs to know every organization that has ever given you money. And it is great to thank people, but if everyone is special, no one is special. Instead of making your final slide into a Guess Who board, consider alternative options to show gratitude. For example, if an undergrad was integral to an experiment, pop up their photo on the results slide and thanks them then. If an advisor was especially helpful, they’ll appreciate a handwritten thank you note more than a pixelated mug shot at the end of your presentation.

Keep your conclusion slide simple. Use the final slide to give your audience ways to learn more. For instance, I try to make a blog post about my presentation that folks can use to find out more information, check out my references, or see my original figures.

Good luck and happy presenting!

At the beginning of my PhD, I had TOO MUCH time to read, but TOO LITTLE focus to know what I needed to read. Now that I have a few experiments behind me and multiple, ongoing analyses and manuscripts in front of me, I have TONS of papers I want to read, but NO time to read them.

To make matters worse, I chase references like a dog chases squirrels. I can sit down with the best intentions of reading through a paper and find myself an hour later down five other rabbit holes with 50 new tabs on my browser all with new papers I certainly won’t ever read.

In an effort to make a consistent practice of winnowing away my ever-growing pile of “really important papers that I definitely want to cite in my dissertation,” I’m starting the #APaperADay challenge. My goal is to read a new paper each workday and write a quick synopsis. I have no roadmap or themes, so reader beware–these paper will be all over the place!

Be sure to check out the whole comment thread for the entire synopsis:

Attention: Dragons are what they eat… kinda.

Here’s my #APaperADay synopsis of @BethanColquhoun @lsweyrich N. Kent & @c_frere ‘s recent paper in @molecology about potentially adaptive shifts in the microbiomes of urban water dragons. 1/7

— A. Z. Andis Arietta (@azandisarietta) November 12, 2019

I love #wilderness, so my #APaperADay summarises @m_dimarco et al.'s assessment in @nature of the impact of human land-use and habitat protection on #extinction risks, in the thread below. 1/6https://t.co/xVM2FjlzAy

— A. Z. Andis Arietta (@azandisarietta) October 21, 2019

Today's #APaperADay is a newly accepted @molecology metropolitan meta-anlysis by @lindsay_s_miles @urbanevol @evoecolab et al that wrangles the explosion of studies to test the effect of urbanization on wildlife. They ask: how do cities help or hinder gene flow (and drift). 1/5

— A. Z. Andis Arietta (@azandisarietta) September 15, 2019

Can bees dance their way to a healthier diet when stuck in a sea of junk food? Nürnberger et al.'s new paper in @molecology suggests so.
Check out my #APaperADay thread below for details. 1/8 #HoneyBee #WaggleDance #Ecology #bees https://t.co/QqvFrzS36z

— A. Z. Andis Arietta (@azandisarietta) August 15, 2019

Has citylife caused adaptive evolution in wood frogs?

To test this, @Jared_Homola et al. (2019) found 4 pairs of frog ponds where one was situated in a rural setting and the other next to a town and then looked for signatures of parallel genetic differentiation. 1/6#APaperADay https://t.co/Rg2b0ICTGk

— A. Z. Andis Arietta (@azandisarietta) July 16, 2019

Pond temps matter a lot to ectotherms like frogs. Mueller et al. (2019)
found that temps experienced at different lifestages impacts frogs’ development and physiology, but the carry-over effect across stages is inconsistent. #APaperADay https://t.co/02AKgI4os7

— A. Z. Andis Arietta (@azandisarietta) July 8, 2019

Pašukonis et al. (2019), found that dart frog dads travel farther than needed when dispersing tadpoles by strapping tiny GPS-transmitters to their backs. #APaperADay https://t.co/G7SGSzbBHR

— A. Z. Andis Arietta (@azandisarietta) July 7, 2019

I’m really impressed with their position. It’s clear that they’ve already thought about the problem of unpaid internships a lot.

I especially want to highlight that I completely empathize with PARC in that these issues are moving targets, especially from the perspective of large, all-volunteer organizations. That PARC is actively working on fixing the problem is to their credit.

I also want to make a strong point that I failed to fully articulate in my last post: the responsibility shouldn’t fall solely on the organizations to fund internship—all of us that appreciate and benefit from the work of those organizations do should feel responsible, too.

The responsibility shouldn’t fall solely on the organizations to fund internship—all of us that appreciate and benefit from the work of those organizations do should feel responsible, too.

I feel really fortunate that I landed a paid internship (shout out to Sitka Conservation Society) right out of undergrad. When that internship rolled into a salaried position, I was already inculcated into the stance that if we couldn’t afford an intern, then we couldn’t offer an internship. But I’ve also served on the Board of Directors for a couple of non-profits and have struggled with the desire to get work done on a slim budget and the temptation to seek willingly free labor from unpaid interns.

The root of the problem is that all of us undervalue the important work of non-profits. If our nonprofits were well funded, this issue would never arise. Unfortunately, environmental organizations receive the least philanthropy of any sector (grouped with animal groups, the sector receives just 2.8% of total charitable contributions annually).

I fear that my first post came across as more of a call-out of NEPARC than a general call-to-action. I’m not a fan of call-out culture, so I hope you will join me in this call-to-action to support PARC. I decided to pony up on my offer to support PARC. I really hope you will make a donation too. They suggest that the best ways to support them are to donate to their non-profit partner, the Amphibian and Reptile Conservancy, or buy some sweet PARC swag.

Or, just take a minute to donate to your favorite environmental non-profit, and feel good that you are helping to end the need for unpaid internships.

Here is the response from PARC, in full:

“Dear Andis,

I’m writing to you on behalf of PARC‘s Executive Committee with regard to your recent blog post: The problem with unpaid pseudo-internships.

First, I’d like to thank you for sharing your perspective and for highlighting actionable steps that PARC must take to ensure equitable and just practices within our organization. I’d also like to apologize for the delayed response; PARC is an all-volunteer organization and it often takes a few days (if not longer) to gather and address feedback from all of the appropriate entities.

We (the National PARC leadership) agree with your views on the issue of unpaid internships in ecology/conservation. This is an issue we have been working to address for the last year. In fact, we have restructured our internships at the national level of PARC (i.e., within the Executive Committee, which oversees the regional and state chapters) to reflect some of the key points addressed in your blog post. In some cases, we have opted to hire contractors rather than creating internships. In other cases, we have opted to provide hourly compensation to interns and to intentionally model the positions in a way that provides the intern with clear learning- and skill-based objectives and opportunities for professional development. With this updated model, we hope that our interns gain as much value from us as we do from them.

As this is a relatively new approach for us, we have not yet developed guidance on this issue for PARC‘s regional and state chapters (again as a volunteer driven organization, these things can take some time). Our extensive discussions were put in place at the national level but never translated into policies and/or guidelines. This, we believe, is a failure on our part. We hope to remedy this shortcoming by taking the following actions:

1 – We will provide time/space for discussion regarding the points you’ve raised in your blog post on our Joint National Steering Committee (which includes regional co-chairs and external partners) and National Diversity, Equity, and Inclusion task team (DEITT) monthly conference calls. Following these discussions, we will work with NE PARC to ensure that appropriate adjustments are made with regard to the social media position.

2 – We are currently developing a best practices document for engaging members and recruiting leaders at the regional and state levels. We will include a section on internships that will provide guidelines for creating equitable and ethical internships.

3 – We will ask the DEITT to provide feedback on our internship guidelines to ensure they reflect PARC‘s goal of providing an equitable platform for our members, partners, and stakeholders to engage in the conservation of amphibians and reptiles.

Thank you again for taking the time to bring this issue to our attention. We are hopeful that in the future, with the assistance of the DEITT, we can be more proactive in addressing these kinds of issues. If you are interested in joining PARC‘s DEI efforts, please consider reaching out to the DEITT co-chairs Neha Savant & David Muñoz (copied on this email). I’ve attached a document that highlights the team’s recent projects/accomplishments.

Best,

Alex Novarro (on behalf of PARC‘s Executive Committee)“

After sleeplessly struggling through my own quals, and afterwards reflecting on the process with friends, I’ve compiled my thoughts/suggestions/tips into this post for those staring down the barrel of their exams.

I’ll admit that I had never heard of qualifying exams until well into the first year of my PhD, and I had absolutely no idea what they entailed or how they transpired. I’m a first generation student, so maybe most folks are better familiarized with the concept. For those who are academically naïve like me, scroll to the end for an overview of the quals process.

Tips for quals:

Prep:

In my experience quals felt a lot like running a marathon through a Ninja Warrior gauntlet, while wearing a T-Rex costume and being chased by a flock of angry mallards.

In other words, you’ll have a lot on your plate. You won’t want to stop for day-to-day chores like paying bills and cooking dinner. Prior to your exams, I suggest trying to preemptively clear as much off your to-do list as possible so that you won’t be distracted nor will you compound your stress load.

For instance, you might pre-pay your bills, do a big grocery run to stock the kitchen with brainy-stimulating snacks, and even pre-prepare freezer-meals to last the duration of your exams. Set your email auto-responder to a plea-ful apology. If you are really savvy, you might even consider giving your partner a few extra backrubs or doing a couple extra chores for your labmates in an effort to bank some reciprocity points that you can cash in for favors during your exams.

Studying:

In some institutions, it is common for committees to give reading lists of critical papers prior to the exams. This isn’t common in my program, but nonetheless, there were about a hundred papers I knew that I would want to have at hand and a few hundred more that I wanted to be able to recall quickly.

You don’t want to be spending precious minutes of exam time searching through databases for the right paper. So, make sure that you have a good reference software onboard and consider blocking out a chunk of time each day leading up to your exam organizing key words and folders you expect to be relevant to your topic.

It is also a huge timesaver to have a reference software that integrates with your word processing application. When you are in the zone, there is nothing more frustrating than breaking the flow to stop and figure out your parenthetical citations. (I use Paperpile and strongly advocate for it).

Do some mental calisthenics to prep your expectations:

Great, now you’ve read and cataloged the relevant literature, banked some favors, and stocked your fridge—you are prepared for your exams! …right?! No matter how confident your feel about your exams, there is bound to be a point during your writing when you may come to one or more of the following realizations:

• There is a lot more literature on your topic than you thought
• You missed a whole boatload of important papers
• Your points are not original
• You know nothing
• Everyone else is smarter
• You will certainly fail your exams
• It might not be too late to go into law school
• Maybe being a barista for the rest of your life wouldn’t be so bad
• Perhaps it would be preferable to live in a cave in the woods forever
• Etc

I call this the “Dark Times”–when the gravity of your examination flattens any modicum of confidence and extinguishes all motivation but the desire to eat Nutella by the spoonful and rewatch the entire first season of the Office in bed.

I’m sure there are countless self-help blogs littered with suggestions for coping with stress, anxiety, and melancholy. But, for now, I share my preferred mind-hack that worked well for me during quals—a mind-hack that comes courtesy of the Stoic philosopher Seneca: “premeditation malorum” (The literal translation is “meditate on future evils.”) The idea is to think about all of the bad things that could happen and all poorer potential alternatives to your situation, thereby vaccinating yourself from the emotional distress if those bad scenarios come to fruition. For example, a few critical remarks from a committee member during orals might be emotionally difficult to take in, but if you’ve been imagining an even worse alternative wherein your exams ends with your committee hurling cabbage and screaming insults, a few critical remarks are no big deal in contrast.

Meditating on poorer alternatives puts things in perspective. After all, when you really think about it, the exam process is just sitting down to do nothing but think about one topic  for a few entirely undistracted weeks (presumably, since this is the topic you’ve chosen to dedicate your life to, this is probably a topic that you are really interested in). If you imagine that you could have spent those weeks endlessly swimming away from hungry sharks, or listening to Dave Mathews Band on repeat, or continuously flossing… the quals seems like down-right eudemonia in contrast.

Another mind-hack is to remind yourself of the purpose of qualifying exams—to push you into the shady corners of your own knowledge and the rarified boundaries of your field. If you start to feel like you don’t know anything, that’s a good sign that you are uncovering a gap in your own knowledge, a gap you can spend the next few years filling. Or, maybe you’ve discovered an unexplored horizon in the field that you can spend the next few years illuminating.

Try to keep in mind that the purpose of quals is to show you what you do not know. Quals are kind of the antidote to academic Dunning-Kruger effect.

It might also be helpful to remember that your committee would never have let you initiate the exam process if they did not expect you to succeed. If you were able to answer all of their questions with ease it would be less of an indication of your expertise than an indication that they failed to their design questions well enough

Plan a perspective parallax:

I hit two severe “Dark Times” during my written exams. In both cases, the thing that halted my unchecked plummet into existential duress were unplanned conversations with peers who had already gone through the exam process. It came as such a relief to hear that they had shared many of the same feelings during their exams. It was heartening to hear that they too discovered papers they missed, to hear that they questioned their entire graduate prospects, and to know that despite it all, they came through the process successfully. Sometime that simple perspective-check from an empathetic peer is a powerful tool to steer your thoughts away from the perigee of despair.

Celebrate:

After my exams, I went to the bar to celebrate. Over beers I had to confess that I kinda liked the process. Yes, it was stressful. Yes, I would not want to revisit the Dark Times. But, I am also incredibly grateful for the chance to spend four solid week thinking about these topics. I recon I learned as much in those weeks as I did in the previous three semesters. The task of explaining a topic to someone else is an entirely different ballgame from convincing yourself that you understand a topic, and doing so solidifies your understanding unequivocally.

At the end of the process, don’t forget to celebrate. More importantly, remember to celebrate your peers going through the same process. After all, we’re all inhabiting this Ivory Tower together, so let’s make it a party!

What are qualifying exams?

Qualifying exams (a.k.a. Quals, Preliminary Exams, Candidacy Exams, etc) act like a turnstile on the path toward one’s dissertation. The purpose of these exams is to demonstrate that you possess the requisite substrate of knowledge in your field. A passing mark on qualifying exams is an endorsement that you are ready and able to define a novel research topic and carry it through into a dissertation. With the exams successfully completed, a PhD student is said to advance to candidacy for a PhD degree, and becomes a PhD candidate.

The structure of the exams is different at every institution, but most seem to include written and oral components. Other than that, the form of exams seems to be widely diverse and variable even between programs in the same institution. So, rather than analyze all of the many exams structure, I’ll quickly explain how exams worked for my program at Yale School of Forestry and Environmental Studies (FES).

At FES, exams tend to center around a student’s prospectus, a document outlining a student’s research plan for their dissertation. For the first year or two of one’s PhD studies, the prospectus is a living document that helps in formulating research questions and outline a plan for answering them. Concurrently with crafting a prospected research plan, students are expected to begin putting together a committee. This is the committee that will preside over the qualifying exam and usually is also the group that will eventually assess the final dissertation. In my case, I have four committee members, all of whom are experts in various aspects of my proposed research.

At FES, the convention is to draft an extremely thorough and extensive prospectus with a full introduction, methods, analysis, and preliminary results sections, most of which will be imported into future publication and the dissertation. Over the months leading up to the exams, the prospectus is floated around to all of the committee both to refine it, but also to give the committee a sense of what the student might be missing.

I planned to take my exams at the end of my third semester (which was pushed back to the beginning of my fourth semester), which is a littler earlier than average for the program. Most students qualify in their fifth or sixth semester.

In FES, the exam is structured in two parts. First, the committee delivers 2-3 essay question that the student is allotted 2 weeks to return answers. After a brief break, the committee convenes to for the oral component. In the oral exam, the student gives a presentation of the prospectus and then the committee gets about 2 hours to ask questions about the prospectus and the student’s answers to the written component. Once the committee has sated their questions, they go into close-door deliberations, and within a 15-30 minutes wither accept, accept with reservations, or deny the student’s advancement candidacy.

The whole process is unequivocally nerve-wracking. After all, you just spent months developing a research plan, then invited a panel of the top experts in that field to grill you on your short-comings over the course of 3-4 weeks. Although it tends to reduce even the most confident PhD student into a sobbing, melancholic, mess of existential-crisis, I unrepentantly appreciated the opportunity and hopefully came out with some tips for other soon-to-be PhD candidates.